資源描述:
《基于關(guān)系數(shù)據(jù)庫(kù)本體自動(dòng)的構(gòu)建地地研究》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、摘要本體作為一種能在語(yǔ)義和知識(shí)層次上描述信息系統(tǒng)的概念模型建模工具,在知識(shí)工程領(lǐng)域得到了廣泛的應(yīng)用,有效的解決了知識(shí)工程發(fā)展過(guò)程中的兩大障礙:知識(shí)的重用和知識(shí)的共享。然而本體的手工構(gòu)建是一項(xiàng)復(fù)雜且容易出錯(cuò)的工作,導(dǎo)致了本體構(gòu)建周期長(zhǎng),成本高。論文研究旨在找出一種利用已有的關(guān)系數(shù)據(jù)庫(kù)資源自動(dòng)構(gòu)建本體的方法,提高本體構(gòu)建的效率。在研究和分析了本體的相關(guān)概念以及現(xiàn)有的本體構(gòu)建方法的之后,將現(xiàn)有方法中數(shù)據(jù)表與概念的映射擴(kuò)展為數(shù)據(jù)表與原始本體的映射。在此基礎(chǔ)之上,提出了一種新的基于關(guān)系數(shù)據(jù)庫(kù)的本體構(gòu)建方法,將本體構(gòu)建過(guò)程分為原始本體抽取和原始本體整合兩個(gè)主要部分。然后分別就這兩個(gè)部分
2、進(jìn)行了細(xì)致的研究。在原始本體抽取的過(guò)程中,本文運(yùn)用了形式概念分析的方法對(duì)數(shù)據(jù)表元組數(shù)據(jù)進(jìn)行分析,從中發(fā)現(xiàn)隱含的概念及其關(guān)系,進(jìn)而生成原始本體。在原始本體整合的過(guò)程中,本文首先采用基于概念相似度聚類的方法對(duì)原始本體的頂層概念進(jìn)行集成,生成概念層次結(jié)構(gòu),然后再對(duì)下層概念的層次結(jié)構(gòu)進(jìn)行調(diào)整,消除語(yǔ)義分歧,最終完成原始本體的整合工作。最后,基于上述方法設(shè)計(jì)并實(shí)現(xiàn)了一個(gè)基于關(guān)系數(shù)據(jù)庫(kù)資源的本體自動(dòng)構(gòu)建系統(tǒng)。并詳細(xì)介紹了系統(tǒng)的各功能模塊及其運(yùn)行流程。關(guān)鍵詞知識(shí)工程,本體,本體抽取,本體整合,形式概念分析ABSTRACTAsamodelingtoolwhichcoulddescribec
3、onceptsonbothofsemanticandknowledgelayers,ontologyhasbeenwidelyusedinknowledgeengineeringfield.Itsolvedthetwoprobleminthedevelopingofknowledgeengineering:knowledgereusingandknowledgesharing.Unfortunately,buildingahigh—qualityontologybyhandsisacomplicatedandmistakablework,SOthattheperiodofb
4、uildingaontologyislongandthecostishigh.Thepaperistofindamethodforautomaticontologybuilding,whichcanusethedatabaseresources,toimprovetheeffigencyofontologybuilding.Afterresearchingontologyandit’Srelatedconceptsandanalysisingexistedontologybuildmethod,thepaperbringoutanewontologybuildmethodw
5、hichimprovedthedatatable—to—one—conceptwayinexistedmethodtothedatatable-to-semi—finished—ontologyway.Basedonthenewmethod,theprocessofontologybuildingisdividedintotwomainparts:semi—finishedontologyextractionandsemi—finishedontologymerging.Thenthepaperresearchedthetwopartsdetailed.Inthesemi—
6、finishedontologyextraction,F(xiàn)ormalConceptAnalysisisusedtoanalysisthetuplesofdatatableSOthatthehiddenconceptsinthedatatableandtheirrelationcanbefound.Inthesemi—finishedontologymerging,ahierarchicalclustermethodbasedonconceptsimilarityisusedformergingtop--levelconceptsofsemi--finishedontology
7、togenerateahierarchicalconceptstructure,thenthelower-levelconceptsisadjustedforeliminatingsemanticconflictstofinishthesemi—finishedontologymerging.Finally,basedonthepaperhadalreadydone,aOntologyGeneraterBasedonDatabase(OGBD)wasestablished.Thenthefunctionmodule